The Edge AI Sensing Platform Discovery Kit brings advanced sensing capabilities and AI computation to the edge. In its first release, it features a SONY IMX335LLN-D image sensor, powered by STM32N6 MCU, which includes the Neural-Art Accelerator™ an ST specific Neural Processing Unit. This combination enables fast deployment, near real-time inferencing, and local event handling, eliminating reliance on cloud interactions.

1. Camera Module:

• SONY 5Mpx IMX335 HDR sensor: The IMX335LLN-D is a diagonal 6.54 mm (Type 1 / 2.8) CMOS active pixel type solid-state image sensor with asquare pixel array and 5.14 M effective pixels. This chip operates with analog 2.9 V, digital 1.2 V, and interface 1.8 V triple power supply, and has low power consumption. High sensitivity, low dark current and no smear are achieved.
• M12 mount with default lens

2. Compute Module:

• STM32N6 contains a  Neural-Art Accelerator™, an ST-developed Neural Processing Unit. IP and peripheral include a MIPI CSI camera, a machine-vision image signal processor (ISP), an H.264 video encoder, and a Gigabit Ethernet controller with Time-Sensitive Networking (TSN) endpoint support.
• 64MB of external NAND Flash
• 32MB of external SDRAM
• Onboard sensors

3.I/O Module:

• USB-C connector
• Expansion header

Software Ecosystem Tools:  STM32Cube.AI 

Documentation for setup and integration: here

*Power adapter and connectivity cables: (sold separately)

Hardware

• Machine vision pipeline: High performance and CPU offload
• Low power high sensitivity image sensor: Captures high-quality images for edge AI tasks.
• Multiple Connectivity Options: USART, USB and (WIFI coming soon)

Software

• Support for multiple object detection models: build, train, and deploy custom models using computer vision models (YOLO, etc).
• On-device AI processing: run AI models directly on devices for real-time responsiveness and reduced latency.